"kolla-ansible destroy" failed when the deployment node is a standalone node

[liumengke]

When the deployment node is separated from the control node, only the kolla-ansible tool is installed on the deployment node, and the docker service is not installed. At this time, the cloud platform can be deployed normally, but when the command "kolla-ansible destroy" is used to clean up the environment, an error about docker command is reported on the deployment node.

It failed when task "Ensure the docker service is running" is executed, and the error message is as below:

----------------
TASK [destroy : include_tasks] ***************************************************************************************************************************
included: /usr/local/share/kolla-ansible/ansible/roles/destroy/tasks/validate_docker_execute.yml for Untitled0, Untitled1, localhost

TASK [destroy : Ensure the docker service is running] ****************************************************************************************************
fatal: [localhost]: FAILED! => {"changed": true, "msg": "non-zero return code", "rc": 1, "stderr": "User root can't seem to run Docker commands. Verify product documentation to allow user to execute docker commands\n", "stderr_lines": ["User root can't seem to run Docker commands. Verify product documentation to allow user to execute docker commands"], "stdout": "", "stdout_lines": []}
changed: [Untitled1]
changed: [Untitled0]

TASK [destroy : include_tasks] ***************************************************************************************************************************
included: /usr/local/share/kolla-ansible/ansible/roles/destroy/tasks/cleanup_containers.yml for Untitled0, Untitled1
-----------------

[liumengke (liumk)]

I have commit a code review for this bug. https://review.opendev.org/c/openstack/kolla-ansible/+/867349